digiLab-ai/uncertainty-engine-sdk ? reverse-engineered prompt

Build me a Python SDK for an Uncertainty Engine service that lets me log in with a username and password, connect to different environments like production or dev, and run nodes on the server from Python. I want a simple Client I can create and authenticate, plus a way to pass in a custom environment if needed. It should support the basics out of the box, and also have optional extras for working with data, notebooks, and graph visualisation.

Please include a few clear examples showing how someone would set their credentials with environment variables, create a client, authenticate, run a simple node like adding two numbers, and read the result. If there are auth tokens or cached login details, handle that smoothly and explain what to do if switching environments causes issues. Keep it easy to install with pip and make sure the code and docs feel polished for a pre release SDK. If you need to, look up the current docs online first.